Fix #73 and support immutable dates

[kylekatarnls]

• Allow DateTimeImmutable
• Fix nextOpen() and nextClose() don't return the correct values when called simultaneously #73: do not modify the passed datetime
• Allow multiple calls for nextClose/nextOpen

[freekmurze]

Remove this, it should be in a global gitignore

[kylekatarnls]

Cleaned, it will disappear if you re-open the PR.

[freekmurze]

Unfortunately changing the return type is a breaking change.

[kylekatarnls]

Nop because your PHP requirement: php: ^7.0 implies that any DateTime object implements DateTimeInterface, so unless using reflexion, there is no break, but with reflexion, anything is breaking change.

[freekmurze]

Closing this as changing the return type is a breaking change. We'll revisit this when creating a new major version of the package.

[kylekatarnls]

@freekmurze Nop because your PHP requirement: php: ^7.0 implies that any DateTime object implements DateTimeInterface, so unless using reflexion, there is no break, but with reflexion, anything is a breaking change.

So you should not consider it as a breaking change and could safely implement this in the current major version.

[kylekatarnls]

Then you still have the missing reset($timeRange);, this bug should be fixed in the current version.

[freekmurze]

It is a breaking change.

Consider this code:

class A
{
    public function date(): DateTime
    {
        return new DateTime();
    }
}

class B extends A
{
    public function date(): DateTime
    {
        return new DateTime();
    }
}

(new B)->date();

If you change the return type of class A to DateTimeInterface you'll get an error:

Declaration of B::date(): DateTime must be compatible with A::date(): DateTimeInterface

[kylekatarnls]

Still, this could have been done differently, simply closing this PR rather than asking for modifications (or take it partially to get the bug fixes, or tag it to be merged in next major version) seems very counter-productive to me. I recommend you a more encouraging contributing process for your PRs such as discuss first, propose alternatives, then decide when every participant finished to consider possible solutions.

[freekmurze]

Thank you for your work around this. I'm sorry that my comments seems counter-productive to you.

Feel free to either rework this PR (in that case I'll reopen it when you push changes) or submit a new PR altogether.